Skip to main content
ARS Home » Research » Publications at this Location » Publication #134371
Title: VISIBLE AND NIR SPECTROSCOPY COUPLED TO IMAGING FOR FOOD PROCESSING AUTOMATION

Author
item Windham, William
item Lawrence, Kurt
item PARK, BOSOON - UGA
item Smith, Douglas

Submitted to: Near Infrared Spectroscopy International Conference Proceedings
Publication Type: Abstract Only
Publication Acceptance Date: 5/6/2002
Publication Date: 5/6/2002
Citation: Windham, W.R., Lawrence, K.C., Park, B., Smith, D.P. 2002. Visible and nir spectroscopy coupled to imaging for food processing automation. [abstract] Near Infrared Spectroscopy International Conference Proceedings.

Interpretive Summary:

Technical Abstract: USDA ARS initiated in 2000 at the request of the Food Safety Inspection Service (FSIS) a research project to investigate the potential of imaging technology for detecting fecal and ingesta contamination on poultry carcasses. Accidental carcass contamination with feces during processing is thought to be the primary route by which pathogens enter the food chain. To accomplish this goal, a logical progression of researchable objectives was developed beginning first with principles and culminating in the validation of an on-line system prototype. First, using a spectrometer on static samples, key wavelengths were identified where feces and ingesta can be distinguished from uncontaminated skin. Second, uncontaminated and fecal contaminated poultry carcasses were imaged with a hyperspectral imaging system. Algorithms were then developed which combine the key wavelengths identified and various image processing techniques to accurately sense any fecal or ingesta contamination. Next an iterative cycle was implemented whereby additional factors that could potentially affect system results were incorporated into both the spectroscopic and hyperspectral imaging techniques above. Hyperspectral image processing algorithms, and histogram stretching, were effective on the identification of fecal and ingesta contamination. Third, once the algorithm and system parameters were determined on the static samples, a multispectral imaging camera was prototyped. This paper presents the logical progression of research that results in a method and imaging system which can be used effectively for detecting feces and ingesta on poultry carcasses It further, demonstrates potential application for on-line processing of poultry for safety inspection.